Design Simulation and Performance Analysis of Real Time Facial Features Monitoring for Drowsiness Detection Using Support Vector Machine?

Abstract: This article describes an effective method for performing drowsiness testing through three well-defined stages. These three stages are using Viola Jones for facial characterization, eye tracking and yawn detection. Once the face is detected, the system illumination is left unchanged by separately segmenting the skin portion and considering only the color components to reject most of the non-face image background based on the skin color. Eye tracking and yawn detection are accomplished by correlation coefficient template matching. Feature vectors from each of the above stages are connected, and a binary linear support vector machine classifier is used to classify successive frames into fatigue and non-fatigue states, and alerts the former if it is above a threshold time. Extensive real-time experiments have proven that the proposed method is very effective in finding sleepiness and warning

Keywords: Drowsiness, Alertness, FFT, ANN, SVM , Face Detection; Eye's State; Drowsiness